MPICH collectives use MPIR_Localcopy to move data from src to destination
buffer for same task. MPIR_Localcopy can't handle GPU buffers. A change
in MPIR_Localcopy would affect all common code. This change is to handle
the checking of GPU buffers in the PAMID collectives layer and allocate
host buffer and copy data from GPU buffer to Host buffer and vice versa so
MPIR_Localcopy would work w/o issues.
This is not performance optimized code.

(ibm) D202834

Signed-off-by: Su Huang <suhuang@us.ibm.com>

Track this library separately so as not to impose its requirements
when building other components like ROMIO or Hydra.

No reviewer.

Commit [7dfe2840

] broke builds with error checking disabled. We fix this
by moving the guards to ensure the fn_exit and fn_fail labels are always
present and only the error checking code is preprocessed out.

Fixes #2225

Signed-off-by: Rajeev Thakur <thakur@mcs.anl.gov>

This code was using uint32_t as a fixed size variable to simplify the
code, but we don't currently require C99. The patch removes that usage
and does everything with int's and more math.

Fixes #2245

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

This is another function that is no longer used as it was replaced by a
static function instead. This also removes an internal test that is no
longer useful for this code.

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

Originally this function was implemented in the CH3 layer, but it was
moved up to the MPI layer as an MPIR function. The old MPID version
doesn't need to exist anymore.

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

The Myrinet MX network module, which had a life cyle from 1.1 till
3.1.2, has now been deprecated.

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

No reviewer.

File ch3u_rma_oplist.c contains functions that make progress
on RMA on the origin side. Here we change the name of file to
a more suitable one.

No reviewer.

Removes additional whitespace in testlist files to make them easier
to manipulate with tools like sed.

No reviewer.

Put tests that check if MPI correctly detects aliased buffers in
collective operations into the errors section of the testsuite.

Fixes #2211

Signed-off-by: Wesley Bland <wbland@anl.gov>

No reviewer

Sangmin Seo authored Mar 05, 2015 and Rob Latham committed Mar 05, 2015

Although MPIDI_CH3I_progress_blocked is a variable only used in CH3,
it was referenced in the ROMIO glue code. This caused a build problem
when pamid is used as a device. This patch removed the reference to
MPIDI_CH3I_progress_blocked, but it degrades the efficiency of
MPIR_Ext_cs_yield_allfunc_if_progress_blocked() since we do not have
a way to check if the progress engine is blocked for now (related to
ticket #2202). For a better solution for ticket #2202, we need to
fix a wait function of the extended generalized request.

Fixes #2242

Signed-off-by: Rob Latham <robl@mcs.anl.gov>

comm_idup was caught failing on mpich-portals4 with configuration
"intel,strict,ib”. It was not fully tested on portals4 because portals
was added after comm_dup patch. On other platforms comm_idup seems to be
OK.

Ticket #2243

No reviewer

Since MPI-3.1 has not been voted this time.

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

Signed-off-by: Pavan Balaji <balaji@anl.gov>

This reverts commit 077346f6

.

Signed-off-by: Wesley Bland <wbland@anl.gov>

This reverts commit c2ea6afc

.

Signed-off-by: Wesley Bland <wbland@anl.gov>

This reverts commit 38df8d2a

.

Signed-off-by: Wesley Bland <wbland@anl.gov>

No reviewer

Xin Zhao authored Mar 03, 2015 and Pavan Balaji committed Mar 04, 2015

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 03, 2015 and Pavan Balaji committed Mar 04, 2015

In MPI standard, predefined datatype is called as basic type.
It is better to make the name same with the standard in the
code.

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 03, 2015 and Pavan Balaji committed Mar 04, 2015

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 03, 2015 and Pavan Balaji committed Mar 04, 2015

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Feb 24, 2015 and Pavan Balaji committed Mar 04, 2015

The implementations of sendNoncontig for intra-node communication in
Nemesis and inter-node communication in network modules (except for
TCP and SCIF) assume that req->dev.segment_first is zero and
req->dev.segment_size is the size of data, which is not always true.
If we stream an RMA operation and issue partial of derived data,
req->dev.segment_first specifies the current starting location of the data
and req->dev.segment_size specifies the current ending location of the data.
Also, the data size should be (req->dev.segment_size - req->dev.segment_first).
This patch corrects this issue in Nemesis and network modules.

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 03, 2015 and Pavan Balaji committed Mar 04, 2015

The original implementation of ACC/GACC on SHM first
allocates a temporary buffer which has the same data
layout as the target data, copies the entire origin
data to that temporary buffer, and then performs the
ACC computation between the temporary buffer and the
target buffer. The temporary buffer can use potentially
large amount of memory.

This patch fixes this issue as follows: (1) SHM ACC/GACC
routines directly call do_accumulate_op() function, which
requires the origin data to be in a 'packed manner';
(2) if the origin data is basic type, we directly perform
do_accumulate_op() between origin buffer and target buffer;
if the origin data is derived, we stream the origin data
by copying partial of origin data into a packed streaming
buffer and performing do_accumulate_op() between the
streaming buffer and target buffer each time.

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 01, 2015 and Pavan Balaji committed Mar 04, 2015

For queued ACC/GACC data piggybacked with LOCK, we do not
need to allocate the buffer for the entire operation, but
only need to allocate a buffer with stream unit size. This
patch fixes this issue.

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 01, 2015 and Pavan Balaji committed Mar 04, 2015

On target side, we always allocate a SRBuf with 256K, which
equals to the size of stream unit, to receive ACC/GACC data.

Note that in MPIDI_CH3U_Request_load_recv_iov(), for ACC/GACC
operations, since we already use SRBuf to receive the data
at beginning, we will not use another SRBuf here, in order
to avoid one more memory copy.

Also, we pass the stream_offset in the current RMA packet to
the request struct (when receiving is not finished) and
do_accumulate_op function (when receiving is finished).

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 03, 2015 and Pavan Balaji committed Mar 04, 2015

Originally, do_accumulate_op() is used to perform the ACC
computation on target between data from origin side and
data on the target window. It requires that the target side
must first unpack the received origin data into the same data
layout as the target data before calling this function, which
may consume potentially large of memory.

This patch fixes do_accumulate_op() function in the following
aspects:

(1) It requires that the origin data passed to the function
must be "in a packed manner", which means it looks as if all
basic type elements in the origin data is placed one by one.
Note that the origin data is not necessarily contiguous, since
we may use non-contiguous basic type. If the basic type
is contiguous, then the origin data must be contiguous.

(2) It adds a new function argument, stream_offset, which
specifies a starting location in the target data. This allows
the origin data to work with partial of target data with stream
size.

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 01, 2015 and Pavan Balaji committed Mar 04, 2015

This patch adds req types for FOP operation, and calls FOP req handler
after SRBuf is unpacked.

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 01, 2015 and Pavan Balaji committed Mar 04, 2015

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 01, 2015 and Pavan Balaji committed Mar 04, 2015

Add stream_offset area into ACC-related packets and request struct
to remember current stream unit's starting position in the entire
target data.

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Mar 01, 2015 and Pavan Balaji committed Mar 04, 2015

Add a counter in op struct to remember number of stream units
that have already been issued. For example, when the first stream
unit piggybacked with LOCK is issued out, we temporarily stop
issuing the following units. After the origin receives the ACK
from the target, it can continue to issue the following units.
This counter helps avoid issuing the first unit again.

Signed-off-by: Pavan Balaji <balaji@anl.gov>